Dynamic Matrix Factor Models with Covariate-Driven Varying Loadings

A 3.5 MB research artifact by Xinyu Li, last updated in May 2026, proposing a novel matrix factor model where factor loadings evolve as functions of observable covariates. The associated files include code and documentation for a nonparametric estimation procedure combining kernel smoothing with PCA. Empirical applications demonstrate the method's performance on international trade flows and portfolio returns.

Use Cases

• Modeling international trade flow dynamics based on the covariate-driven factor loadings described.
• Analyzing portfolio return time series with structural breaks based on the proposed dynamic loading framework.
• Testing the hypothesis of constant factor loadings using the described bootstrap-based test.
• Estimating latent factors in high-dimensional settings where loadings depend on observable temporal characteristics.

Strengths

• Includes Python code for implementing the proposed nonparametric estimation method.
• Released under a permissive CC-BY-4.0 license, allowing for reuse and modification.
• The methodology is validated through Monte Carlo simulations and two empirical applications.

Limitations

• Description metadata is limited; actual data quality requires manual inspection after download.
• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment.

Provenance

Source

figshare

Collection Method

Likely contains simulated data for Monte Carlo experiments and empirical application data.

Time Range

null

Freshness

Last updated 2026-05-08 16:10:10; freshness should be verified.

Geography

null